Transmission lines on multiple dielectric layers
时间：03-31
整理：3721RD
点击：
I am trying to model coplanar waveguides (CPW) on top of a silicon substrate (~500um) with thin oxide layer (~10 um) using Agilent Momentum. I am representing the CPWs using slot structures (rather than strips) on the metallization layer.
If I simulate the CPW on just the silicon substrate without the oxide layer, Momentum is able to compute the port impedance which matches well with other solvers. However, when I add the thin oxide layer, Momentum is no longer able to compute the impedance and instead defaults it to 50 ohms. I get the following warnings:
- Grounded coplanar portline impedance could not be computed. continuing simulation with Zo = 50 ohms
- Port solver failed to compute propagation constant and characteristic impedance. 50 ohm default will appear in the dataset
I initially thought this is because of meshing problems associated with the thin oxide layer. However, I encounter the same warnings even after increasing the thickness of the oxide layer to 500um so this may just have to do with multiple dielectric layers with no metallization layers in between.
Are there any special tricks in Momentum when running simulations with two or more dielectric layers, in particular with thin dielectric layers on top of thicker layers?
- LS
If I simulate the CPW on just the silicon substrate without the oxide layer, Momentum is able to compute the port impedance which matches well with other solvers. However, when I add the thin oxide layer, Momentum is no longer able to compute the impedance and instead defaults it to 50 ohms. I get the following warnings:
- Grounded coplanar portline impedance could not be computed. continuing simulation with Zo = 50 ohms
- Port solver failed to compute propagation constant and characteristic impedance. 50 ohm default will appear in the dataset
I initially thought this is because of meshing problems associated with the thin oxide layer. However, I encounter the same warnings even after increasing the thickness of the oxide layer to 500um so this may just have to do with multiple dielectric layers with no metallization layers in between.
Are there any special tricks in Momentum when running simulations with two or more dielectric layers, in particular with thin dielectric layers on top of thicker layers?
- LS